eXplainable Machine Learning does complement usual math modeling in identifying salient variables and their possible nonlinear dynamic interactions from data, as also described in:

Muselli and Liberati, IEEE Trans KDE 2002

Ferrari Trecate, Musellli, Liberati and Morari, Automatica 2003

Garatti, Bittanti, Liberati and Maffezzoli, Intelligent Data Analyis, 2007

Grassi, Liberati et al., Frontiers in Oncology, 2019